2019 HSR&D/QUERI National Conference

1016 — Enrollment and retention of men and women in health services research & development trials

Lead/Presenter: Karen Goldstein,  COIN - Durham
All Authors: Goldstein KM (Durham VA HCS, Duke University Division of General Internal Medicine), Duan-Porter W (Minneapolis VA HCS, University of Minnesota Medical School), Alkon A (Durham VA HCS, Duke University Department of Medicine) Olsen MK (Durham VA HCS, Duke University Department of Biostatistics and Bioinformatics) Voils CI (William S Middleton Memorial Veterans Hospital, University of Wisconsin Department of Surgery) Hastings SN (Durham VA HCS, Duke University Department of Medicine, Duke University Center for the Study of Human Aging and Development)

Sex- and gender-specific science is essential to inform patient-centered, evidence-based care. Developing such evidence requires adequate inclusion of both women and men in trials. We sought to describe representation of men and women across stages of study participation in behavioral intervention trials conducted at the Durham VA Health Services Research and Development (HSRandD) Center of Innovation over the last 10 years.

We obtained information about study participation using our common tracking database supplemented by aggregate study data from study teams as needed. We compared the participation of women and men from trial recruitment to study completion. We also calculated the participation to prevalence ratio (PPR) by sex for each trial.

We included eight trials that started recruitment between 2011 and 2014. In total, these trials consented 2,889 participants, 13.5% of whom were women (n = 389). Included trials addressed patient and caregiver dyads, cardiovascular risk factors including obesity, hypertension and smoking cessation, emergency department follow-up, and osteoarthritis of the knee and hip. Only one study purposefully attempted to boost the recruitment of women. Overall, the PPR for women ranged from 0.2 to 4.5, with 7 studies having a PPR over 1, indicating that women participated in these trials at proportions greater than their prevalence in the relevant disease or condition population within the VA. The PPR for men ranged from 0.8 to 1.1. For both women and men participants, retention was best in those studies that used administrative data for final outcome assessment. No studies provided results stratified by sex or were powered to conduct moderation analyses to explore treatment effect by sex.

At a single site, women participated in HSRandD trials at similar or greater rates to men when invited to participate without cross-study efforts to enrich recruitment or retention by sex.

To achieve equitable benefit from VA research for both women and men, investigators should report results by sex and include sex-based analyses when possible. Adding strategic recruitment approaches could further boost the proportion of women in VA trials and enable adequately powered sex-based analyses.